Fabricated metal product manufacturing

Two trades under one code, both working to order. Stamping is the accessible one on paper — 124 of Canada's 181 establishments, 69%, have fewer than twenty employees, and a press shop running customer-owned dies can be bought from a retiring owner. Forging is heavier: hammers, presses, furnaces and heat-treat, selling into aerospace, energy and heavy equipment where the part must be qualified before it is bought. The pre-screen said "forging presses are the business". The test is whether owning them pays, and the one small forger that reports in public says not reliably. SIFCO Industries, an Ohio aerospace and energy forger, had fiscal 2025 net sales of $84.8M, up from $79.6M, and still lost $0.9M from continuing operations, after losing $8.6M the year before; adjusted EBITDA was $5.7M [A]. It ended the year with a customer backlog of $119.2M — about 1.4 years of sales already ordered [A]. A shop that is qualified, full and growing, and only just approaching break-even, is telling you where the value goes: the equipment and the approvals are sunk by the forger, and the price is set by a handful of primes on long-term agreements. The press cannot be redeployed, and the customer knows it. Stamping shares the shape at lower stakes. The die usually belongs to the customer, so the work can be moved by moving the die; the stamper owns the press and the risk of an idle shift. In Ontario and Quebec, where 87% of these shops sit, the customer is often an automotive tier whose volumes the stamper does not control. The ERP software sold to these shops is screened separately. Among the job-shop trades, the press-based ones are the capital-heavy end, with the least ability to reprice.

The pre-screen called hand tools "an import-and-brand category, not a manufacturing entry". The count agrees about who is actually here: 87 of Canada's 128 establishments, 68%, have fewer than ten employees [A]. These are knife makers, edge-tool smiths, specialty woodworking-tool makers and saw-blade shops. Making the thing is reachable — a forge, grinders, heat-treat, a few CNC machines. At that scale it is a craft business and a good one for the person who wants it. The question a screen asks is whether it becomes more than that. The obstacle is the shelf. Stanley Black & Decker's Tools & Outdoor segment — Stanley, DeWalt, Craftsman, Irwin, Lenox — took $13,158M of net sales in 2025 on a 10.7% adjusted segment margin, inside a company whose full-year sales fell 2% to $15.1B, 1% organically, after a cost programme that has removed about $2.1B of annual cost since mid-2022 [A]. Read that as a description of the channel. Mainstream hand tools reach users through a few home-improvement and industrial distributors whose planograms are negotiated with a supplier of that size, in a category where that supplier's own sales are falling and its energy is going into making the same tools cheaper. A new maker cannot buy its way onto that shelf and cannot match the landed cost if it did. What remains is direct sale of a premium tool to an enthusiast — mail order, the maker's own site, a specialist retailer. That works, and Canada has well-known examples, but each one is bounded by the size of its enthusiast audience, and none publishes results, so how far the ceiling sits above a good living was not established. The ERP software sold to metal shops on this branch is screened separately.

The pre-screen marked this a candidate, and the screen did not overturn that — this record does not find a clean kill, and says so. It is filed with the cut that came closest. What is attractive is real. This is a broad trade, not an oligopoly: 2,373 Canadian establishments, 42% with fewer than ten employees and another 42% with ten to forty-nine [A], led by Ontario, Quebec, British Columbia and Alberta. The shops are certified, equipment-rich and owner-run, which is the profile that sells to a successor. And the top of the industry earns real money: ADF Group, a Quebec fabricator of complex structures, reported a 31.6% gross margin on C$339.6M of revenue in fiscal 2025, and when Canam Group was taken private in 2017 the buyers paid C$12.30 a share, a 98.4% premium, for an enterprise value of about C$875M [A]. Informed money has valued this trade highly. The cut that came closest is the quality of the revenue. ADF's next year shows it: revenue fell 24% to C$258.7M, gross margin to 23.1% and net income from C$56.8M to C$26.3M, with the Terrebonne plant on a work-sharing programme, and the company blames US tariffs and the steel price set by US mills [A] — while its backlog reached a record C$561.1M. Work arrives as discrete projects, each won by bid, often at a fixed price struck before the steel is bought; a full order book and an idle shop can coexist for a year. For a small shop the same lumpiness runs through two or three general contractors. A full study should test, for a specific shop in a specific region: gross margin through a cycle, how steel-price risk is shared in its contracts, bonding and working-capital needs, and customer concentration. The ERP software sold to these shops is screened separately on this branch.

The reachable proposition here is not a can line — that is a continuous-process plant selling to a handful of beverage fillers — but a code-stamped tank and pressure-vessel shop: propane and fuel tanks, heat exchangers, process vessels, welded to a registered design for regional energy, agricultural and industrial buyers. It is a real small-business population (303 Canadian establishments, a third of them under ten people) and the welding-code registration that looks like a barrier can be bought with an existing shop. The cut is what the one listed Canadian owner of such shops reports. TerraVest Industries grew sales 50% to $1,371.2M in fiscal 2025 — and its base portfolio grew 1% ($725.6M against $717.4M) [A]. Every other point of growth was purchased: four tank and vessel makers acquired in the year. That is the shape of the industry stated by its best operator: demand for tanks is flat replacement demand, and returns come from buying shops, consolidating steel purchasing and spreading certified designs across plants. An independent entrant owns one shop in a flat market, buys plate at a worse price than the consolidator, and meets that consolidator again as the natural bidder when it wants to sell. The 19% adjusted EBITDA margin TerraVest earns is a portfolio result, not a single-shop one. Software sold to these shops is screened separately.

Hinges, slides, pulls, locks and latches are simple stamped, cast and machined goods, and a small maker can produce a credible line — 57 of Canada's 108 hardware manufacturers have fewer than ten people. The cut is that the manufacturer does not own the customer. Richelieu Hardware, which describes itself as an "importer, manufacturer and distributor of specialty hardware", sold $1.96B in fiscal 2025, $1.74B of it to manufacturers — the cabinet shops, furniture plants and millworkers who actually consume hardware — and only $220.7M to retailers [A]. It has made 100 acquisitions to build that position. A kitchen-cabinet shop buys from one catalogue of over 145,000 items served out of 119 North American centres; it does not open an account with a single-line maker. So a new manufacturer reaches its buyers through the distributor's catalogue, at the distributor's price, beside the distributor's imported equivalent — and the distributor decides which of the two it promotes. The same structure holds at the retail end, where a big-box buyer sets the planogram. Richelieu's 10.9% EBITDA margin is earned on selection and delivery, not on making things, which says where the value in hardware sits. What remains for a domestic maker is custom architectural and specialty work sold on drawings, which is a job shop under another name and is not a distinct path. Software for metal shops is screened separately.

Buying wire and bending, weaving or welding it into nails, mesh, fencing, racks and springs needs modest machinery, and 113 of Canada's 167 plants employ fewer than twenty people. The cut is who sets the price of a wire product: the import quote. Tree Island Steel of Richmond, B.C. — a listed maker of nails, stucco and concrete mesh, fencing and other fabricated wire products — reported 2025 revenue of $161.8M, down from $207.0M, with gross profit of $9.5M [A]. That is a gross margin under 6% before a dollar of overhead; adjusted EBITDA was $3.0M and the year closed at a $5.3M net loss, the second loss running. The company cut its workforce by 27%, withdrew from product lines it called unprofitable, and lost U.S. volume to expanded tariffs on wire products. A nail or a roll of mesh is specified by gauge and coating, so a distributor can substitute an offshore container for a domestic pallet without the end user noticing, and the domestic plant earns only what is left after matching that price on steel it bought at the North American rod price. An entrant would be the smallest buyer of wire selling into that spread. Custom springs are a different business — made to a customer's drawing, in short runs, with some engineering content — and resemble the job shop more than the wire mill; that niche was not tested here. Software for metal shops is screened separately.

The pre-screen called this a candidate and the screen does not overturn that: this record does not find a clean kill. The population is large and small — 2,497 Canadian establishments, 63% with fewer than ten people; in the U.S. 22,081 shops averaging 16 employees — which means thousands of owner-operators, many near retirement, selling businesses a buyer with ordinary resources can finance. Machines hold resale value and the skill is real. What a full study has to test is what the buyer actually acquires. A machine shop owns no product: its revenue is other companies' drawings, re-quoted each time, and usually concentrated in a few accounts held personally by the departing owner. The order book can walk with him. The second pressure is measured. Protolabs' CNC machining revenue grew 17.6% to $243.3M in 2025 [A] on automated quoting and a network of partner shops, and its annual report names "thousands of alternative manufacturing machine shops" among its competitors. The quick-turn, low-volume job — historically the local shop's best-priced work — is the part migrating to a web upload. What stays local is repeat production, large or awkward parts, and repair for nearby industry. A study should price a specific shop against three things: share of revenue in its top three customers, whether the machinists stay, and how much of the work is quick-turn. Defensibility is named as the cut because it is the weakest factor, not because it is proven fatal. Software for these shops is screened separately.

This is the service end of metalworking: a machine shop sends parts out to be hardened, plated, anodised or coated and gets them back in days. The customer base is local and sticky, the work is priced per lot rather than bid against imports, and most of Canada's 584 establishments are small — 64% employ between five and forty-nine people. The cut is what the process costs to own. Bodycote, the largest listed heat treater, earned a 15.7% adjusted operating margin on £727.1M of revenue in 2025 — and spent £77.0M on capital expenditure, 10.6% of revenue, across 136 locations [A]. That is roughly £5M of revenue per plant, each a building full of furnaces, atmospheres and quench systems that burn energy whether or not the racks are full. More telling is what it is doing with its ordinary plants: 31 sites have been declared non-core — described by the company as "older, less efficient and more carbon-intensive" heat treatment serving automotive and general industry — and are being closed, merged or sold, ten of them in France in one transaction [A]. Revenue fell 4%. The general-purpose shop serving local industry is precisely what the scale operator is exiting; the margin sits in aerospace, gas-turbine and specialist processes that require accreditation and years of customer qualification. Plating adds a second weight the heat treater escapes — chemical baths, wastewater permits and the contamination history of any site bought. An entrant can buy a shop, but it buys a furnace-replacement schedule with it.

A residual group: metal valves, ball and roller bearings, and a miscellany code (332999) that holds everything from safes to metal ladders. Bearings are a global scale industry and the miscellany has no common economics; neither was examined. The one niche with a recognisable Canadian path is industrial valves, where Montreal's Velan has long sold into refineries, power stations and navies. That history is the attraction — engineered product, approved-vendor status, a long-lived installed base — and Velan's accounts are the cut. In fiscal 2026 it reported sales of US$296.4M, up 0.4%, a 27.4% gross margin, and operating income of US$1.6M [A]: about half a percent of sales, from a company that already holds every approval an entrant would spend a decade earning. Bookings rose 0.9%. The backlog of US$283.3M is almost a full year of sales, which describes the business model: large project orders, won by tender, built over many months and financed by the maker until delivery. In the same year Velan paid US$143.0M to settle U.S. asbestos claims and sold its French subsidiaries for US$208.2M to fund it — a reminder that an industrial product's liabilities can outlive its margins by decades [A]. A new valve maker would be qualifying for approved-manufacturer lists in order to join a flat, tender-priced market in which the established Canadian name barely breaks even. Software for metal shops is screened separately.